JOURNAL
OF
EXPERIMENTAL
CHILD
PSYCHOLOGY
22, 67-12 (1976)
Reception Training Effects on the Production Modeled Language Constructions
of
Ross VASTA AND MICHAEL TEITELBAUM State University of New York College at Brockport Several previous studies have reported that first-grade-age children increase their usage of prepositional phrases when exposed to novel (inverted) prepositional constructions. The hypothesis was tested that such anomalous increases could be eliminated by discrimination training on the two prepositional forms. One group of first-grade subjects (age 72 months) was trained to respond differentially to sentences containing either familiar or novel prepositional constructions, while a control group received only exposure to the same sentences. As predicted, when later exposed to the modeling of novel constructions, children in the pretrained group did not increase their frequency of familiar prepositional phrases, while those in the control group did.
In 1966, Banduraand Harris reported that the frequency of prepositional phrase usage by second graders could be increased through modeling and reinforcement procedures. Contending that this demonstration failed to deal with acquisition of a language construction, Odom, Liebert, and Hill (1968) attempted to replicate the study using a novel prepositional construction of the form, article-noun-preposition (“The dish was the table on”). Results obtained with this construction provided a curious phenomenon. Second graders exposed to a model who produced sentences containing the novel construction increased only their frequency of familiar, English prepositional phrases. This occurred again in a follow-up, despite a procedure change requiring the child to repeat the modeled sentences before constructing his own. Liebert, Odom, Hill, and Huff (1969) conducted an extended replication of this study to examine the phenomenon developmentally. Although the oldest group which they studied (ninth grade) was successful in acquiring use of the novel construction, the two younger groups (third and first grades) did not produce sentences containing the novel forms. However, only the first-grade subjects increased their production of familiar prepositional constructions. This finding led Vasta and Liebert (1973) to speculate that differential productive performance by these two younger groups might be a result of differential receptive discrimination ability. Specifically, it was hypothesized that the younger group failed to discriminate the inverted prepositional construction as a novel form and, thus, treated sentences Requests for reprints should be sent to Ross Vasta, Department of Psychology, State University College at Brockport, Brockport, N. Y. 14420. 67 Copyright All rights
0 1976 by Academic Press, Inc. of reproduction in any form reserved.
68
VASTA
AND
TEITELHAUM
containing these forms in a manner similar to the subjects of the Bandura and Harris (1966) research. However, the third-grade subjects, although also unable to produce the novel constructions, were capable of discriminating them as a different form and, therefore, did not increase their frequency of familiar prepositional phrase usage. This hypothesis was tested by again exposing subjects at these two age levels to the modeling of novel constructions but testing them only on an auditory discrimination measure. The results of that study, as well as a subsequent replication (Grant, Note I), confirmed the differential discrimination hypothesis, albeit somewhat indirectly. The present research took a more direct tack. A group of first-grade subjects was pretrained in the novel-familiar discrimination prior to the modeling of novel constructions. This group was predicted not to produce the anomalous increase in familiar phrases, in contrast with a second group, which received only exposure to the pretraining sentences. METHOD
Participants Four males and eight females, ages 64-89 months, were drawn from three classes at a campus school serving both a college (middle class) and community (working class) population. Two black children were included, one in each treatment condition. The Experimenter was a white male undergraduate; the Model was a white female undergraduate. Apparatus
and Setting
The data for all sessions of the study were collected in a windowless, well-lighted cubicle, 2 x 3 m, in the school building. The room contained a table and two small chairs. For the Pretraining session, an apparatus was constructed from white posterboard measuring 40 x 50 cm. A red square and a blue square of colored paper, 1.5 x 15 cm, were pasted on the board, parallel to its edges, approximately 8 cm apart. The red was at the left, the blue at the right. A cardboard arrow was mounted on the board between the squares with an envelope clasp which permitted it to swivel from one square to the other. Between the squares and immediately above the arrow, a second envelope clasp protruded upwards, on which round yellow cardboard disks, 5 cm in diameter, could be placed. Fourteen white posterboard cards, 25 x 40 cm, were also constructed. On each was either a red or blue square, identical in size to those pasted on the apparatus. The red squares were positioned at the left of the card, the blue at the right. For the Modeling phase, a Lafayette tape recorder (Model RK-86A) was used to record the session. The microphone was positioned, unobtrusively, to the subject’s left.
RECEPTION
TRAINING
69
Procedure Subject selection. Children in three classes were matched in pairs for sex and age (in months). The two children from six well-matched pairs were then randomly assigned to either Experimental or Control conditions. The meanages for the twogroups were 72.0and 72.3 months, respectively. Pretruining . Each subject from the Experimental group was run before her matched control. The subject was escorted by the Experimenter to the test room where she was seated at the table and instructed how to play the game. First, the Experimenter produced a round cardboard disk on which he printed the subject’s name and placed it on the apparatus, above the arrow. The subject was told that the game would begin with the arrow pointed to her “name.” The Experimenter explained that he was going to say some sentences and after each one the child was to guess whether it was “a red kind” or “a blue kind.” The subject was instructed to make her selection by moving the pointer to either the red or blue square on the apparatus. Blue was arbitrarily selected as the correct color for sentences containing the novel constructions. The sentences each contained only one novel or familiar prepositional construction and were presented in random order. After each correct response, the Experimenter said, “Right,” instructed the subject to return the arrow to her “name,” and presented the next sentence. After each incorrect response, the Experimenter said, “No, let’s try that one again,” and repeated the trial until the subject responded correctly. Visual feedback was also provided in that, after each response by the child, the Experimenter turned up one of the cards which were piled, face down, just beyond the apparatus. He stood it just behind the apparatus, where its square corresponded in color and position to the correct response for that trial. If the response was correct, the card was put aside; if the response was incorrect. it was placed face-down and turned up again following the next response. Training continued until the subject produced six consecutive correct responses, on thejirst attempt. After criterion performance was reached, the child was awarded her “name” disk and returned to the class. A mean of 16.2 trials was required for subjects in this group. The subject’s yoked control was subsequently run in a similar procedure, but with feedback eliminated. The controls were told that they would be apprised, at the end of the game, of how well they had performed. The cards were not included in the control procedure. Each control was exposed to the same sentences as her experimental counterpart, receiving the same number of trials. Sentences were repeated where such had been the case for the experimental subject, but at the end of the other trials. Following completion, the subject was told she had done well and was awarded her disk. Modeling.The modeling session took place 2-4 days following
70
VASTA AND TEITELBAI!M
pretraining. Each subject was again escorted by the Experimenter to the test room where the Model was already present. The subject was introduced and told that the Experimenter was interested in how people make up sentences. He explained that he was going to say some words and after each one the object was to produce any sentence using that word. The Model and subject were given a sample sentence to ensure that the instructions were clear. The subject was then “randomly” selected to go first and the Experimenter proceeded to present IO common nouns to which the subject responded with sentences. No feedback was delivered during this baseline phase. After the tenth sentence, the Experimenter explained that he preferred some sentences to others and that, thereafter. when he heard one he liked, he would say as much. He then presented five other nouns to the Model, who produced a predetermined sentence following each. Four of the sentences contained a novel construction which the Experimenter met with “Good, I like that one” or a similar phrase. A random fifth sentence contained a familiar prepositional construction to which the Experimenter made no remark. The Model and subject alternated in blocks of five sentences until the subject had produced 20 additional sentences. She was then thanked and taken back to her class. RESULTS The data of interest were the percentages of sentences containing at least one preposition-article-noun construction. A 2 x 2 analysis of variance was conducted with Treatment groups (Pretrained and Control) as a between-group factor and Phases (Baseline and Modeling) as a withingroup factor. A main effect was significant for the Phases factor, F( I, 10) = 11.63, p < .Ol, as well as for the Treatment x Phases interaction, F(l,lO) = 11.63, p < .Ol. Internal analyses revealed that subjects in the Control group significantly increased their frequency of familiar prepositional phrase usage, t(5) = 4.07, p < .Ol, while the Experimental group did not. All six subjects in the Control group increased their usage of familiar prepositional phrases (X = +26.7%). Two of the Pretrained subjects showed an increase, one remained the same. and three showed a decrease (x = 0.0%). Fisher’s exact test verified that the number of subjects who increased their familiar prepositional phrase frequency was significantly associated with their treatment condition, p = .0303. DISCUSSION The results confirm the initial predictions. Despite exposure to a number of sentences containing novel or familiar prepositional constructions and instructions to respond differentially, all subjects in the Control condition once again increased their frequency of familiar prepositional phrase usage
RECEPTION
TRAINING
71
when exposed to the modeling of novel prepositional forms. This demonstration, thus, constitutes the fourth time this phenomenon has been reported. In contrast, for the first time a group of subjects at this age level, who were exposed to the modeling of novel forms, did not increase their familiar phrase usage, presumably as a result of their earlier training in receptive discrimination skills. The data, thus, further support the differentially discrimination hypothesis offered previously (Vasta & Liebert, 1973).
The import if these results goes beyond this particular phenomenon, however. The demonstration that training in receptive skills can facilitate productive performance adds to a growing literature on this topic (e. g., Guess, 1969; Guess & Baer, 1973; Harrelson, Note 2; Mann & Baer, 1971; Vasta, Whitehurst, & Borkowski, 1974). In addition, these data may be compared with evidence relating syntax discrimination to direct imitation. For example, Bohannon (1975) has reported that first-grade children who, without training, could discriminate normal from scrambled syntax, imitated sentences better than did nondiscriminators. In that study, the dependent measure involved instructed exact copying. The present study, however, may have greater relevance for a related dodeling process, selective imitation. Whitehurst and Vasta (1975) have offered a theoretical account of language acquisition, suggesting that selective imitation, i. e., imitation of the form but not specific content of modeled language (e. g., Bandura & Harris, 1966), might be a process through which forms, initially acquired receptively, can be introduced into the productive mode. The present data are consistent with that suggestion in that they demonstrate how receptive facility with a given form can influence the production (or, in this case, nonproduction) of those forms in a modeling situation. REFERENCES Bandura,
A., & Harris, M. B. Modification of syntactic style. Journal of Experimental Child Psychology, 1%6. 4,341-352. Bohannon, J. N. The relationship between syntax discrimination and sentence imitation in children. Child Development. 1975, 46, 444-45 I. Guess, D. A functional analysis of receptive language and productive speech: Acquisition of the plural morpheme. Journal of Applied Behnvior Anulysis, 1969. 2, 55-64. Guess, D.. & Baer. D. M. An analysis of individual differences in generalization between receptive and productive language in retarded children. Jownal oj’App/ied Behavior Analysis. 1973. 6, 311-329. Liebert, R. M., Odom, R. D.. Hill, J. H.. & Huff, R. L. The effects of age and rule familiarity on the production of modeled language constructions. Developmental Psychology. 1969. 1, 108- 112. Mann, R. A., & Baer, D. M. The effects of receptive language training on articulation. Journal of Applied Behavior Analysis, 1971, 4, 291-298.
72
VASTA
AND
TEITELBAUM
Odom.
R. D., Liebert. R. M., & Hill. J. H. The effects of modeling cues. reward, and attentional set on the production of grammatical and ungrammatical syntactic constructions. Jollrnul yf‘Experimentrr/ Child Psychology. 1968. 6, 13 I- 140. Vasta. R.. & Liebert. R. M. Auditory discrimination of novel prepositional constructions as a function of age and syntactic background. Developmental Psychology. 1973. 9, 79-82. Vasta. R.. Whitehurst, G. J., & Borkowski. 0. Accuracy of imitative production and discrimination: Effects of contingent feedback. Perceptrral and Motor skills. 1974. 39, 483-490. Whitehurst. G. J.. & Vasta. R. Is language acquired through imitation’? Jorrrnal qf Psycholinguistic Research, 1975, 4. 37-59.
REFERENCE
NOTES
I. Grant, K. E. Auditory discrimination of nolsel prepositional constructions Unpublished manuscript. 1975. 2. Harrelson, A. Effects of producti\se speech training on receptive language. Master’s thesis, University of Kansas. 1969. RECEIVED:
July 2, 1975: REVISED:
December
17. 1975.
in children. Unpublished